Open Access

Table 1

Classification performance of the models in this work.

Classnames ATAT (LC + MD) ATAT (LC + MD + Features) RF (MD + Features)



Precision Recall F1 score Precision Recall F1 score Precision Recall F1 score
CART 72.8 ± 2.9 40.3 ± 3.8 51.7 ± 2.9 75.3 ± 2.5 40.0 ± 4.6 52.0 ± 3.3 59.2 ± 0.4 56.2 ± 0.6 57.6 ± 0.5
Iax 59.8 ± 1.3 70.1 ± 2.2 64.5 ± 1.2 59.8 ± 2.1 65.1 ± 5.4 62.2 ± 1.9 57.6 ± 0.5 55.9 ± 0.6 56.8 ± 0.5
91bg 89.4 ± 0.4 92.4 ± 1.1 90.9 ± 0.4 88.8 ± 2.2 92.5 ± 1.9 90.5 ± 0.6 75.2 ± 0.4 90.2 ± 0.2 82.0 ± 0.2
Ia 75.7 ± 1.6 84.0 ± 1.7 79.6 ± 0.6 76.3 ± 1.2 81.4 ± 1.7 78.8 ± 0.7 61.4 ± 0.4 76.7 ± 0.2 68.2 ± 0.3
Ib/c 53.8 ± 3.1 63.9 ± 1.9 58.3 ± 1.3 50.0 ± 3.8 65.8 ± 3.4 56.6 ± 1.2 58.0 ± 0.3 39.6 ± 0.4 47.1 ± 0.2
II 65.8 ± 2.3 65.2 ± 1.0 65.5 ± 1.3 63.9 ± 3.5 66.4 ± 2.8 65.0 ± 1.3 66.8 ± 0.6 42.7 ± 0.5 52.1 ± 0.5
SN-like/Other 66.7 ± 1.5 62.9 ± 2.4 64.7 ± 1.7 64.3 ± 2.2 60.5 ± 2.9 62.3 ± 1.5 59.0 ± 0.5 54.1 ± 0.8 56.5 ± 0.6
SLSN 89.0 ± 1.1 95.3 ± 0.3 92.0 ± 0.5 89.6 ± 0.9 95.4 ± 0.4 92.4 ± 0.4 90.3 ± 0.1 90.0 ± 0.1 90.2 ± 0.1
PISN 93.1 ± 1.1 97.5 ± 0.5 95.2 ± 0.4 95.9 ± 0.4 96.7 ± 0.9 96.3 ± 0.4 85.6 ± 0.1 96.7 ± 0.1 90.8 ± 0.0
TDE 77.2 ± 2.9 92.7 ± 0.8 84.2 ± 1.6 79.0 ± 4.9 92.5 ± 1.0 85.1 ± 2.6 83.2 ± 0.4 76.8 ± 0.3 79.9 ± 0.2
ILOT 89.6 ± 0.8 85.9 ± 2.9 87.7 ± 1.1 92.1 ± 0.9 84.0 ± 3.1 87.8 ± 1.3 76.3 ± 0.3 93.6 ± 0.2 84.1 ± 0.2
KN 97.7 ± 0.3 71.5 ± 2.6 82.6 ± 1.7 97.1 ± 0.4 77.1 ± 2.5 85.9 ± 1.4 86.8 ± 0.2 90.3 ± 0.1 88.5 ± 0.1
M-dwarf Flare 98.9 ± 0.4 67.5 ± 1.2 80.2 ± 0.8 99.1 ± 0.3 70.4 ± 1.9 82.3 ± 1.3 95.0 ± 0.3 79.4 ± 0.3 86.5 ± 0.3
uLens 85.8 ± 1.8 95.3 ± 0.9 90.3 ± 0.8 86.8 ± 1.7 95.6 ± 0.7 91.0 ± 0.7 96.9 ± 0.4 82.8 ± 0.2 89.3 ± 0.3
Dwarf Novae 87.9 ± 1.8 86.3 ± 0.9 87.1 ± 0.7 86.2 ± 1.8 92.0 ± 0.9 89.0 ± 0.9 78.5 ± 0.2 82.9 ± 0.3 80.6 ± 0.2
AGN 99.8 ± 0.0 100.0 ± 0.0 99.9 ± 0.0 99.7 ± 0.1 100.0 ± 0.0 99.8 ± 0.1 95.4 ± 0.4 99.9 ± 0.1 97.6 ± 0.2
Delta Scuti 92.3 ± 0.7 95.0 ± 0.4 93.6 ± 0.3 98.7 ± 0.3 99.5 ± 0.1 99.1 ± 0.1 90.8 ± 0.3 98.9 ± 0.0 94.7 ± 0.2
RR Lyrae 93.5 ± 1.0 96.0 ± 0.9 94.7 ± 0.3 99.5 ± 0.2 99.1 ± 0.2 99.3 ± 0.1 91.6 ± 0.4 98.9 ± 0.1 95.1 ± 0.2
Cepheid 96.0 ± 0.8 97.9 ± 0.3 97.0 ± 0.4 99.2 ± 0.3 99.5 ± 0.1 99.3 ± 0.1 92.6 ± 0.5 98.9 ± 0.1 95.6 ± 0.3
EB 87.0 ± 0.3 98.8 ± 0.3 92.5 ± 0.1 90.4 ± 1.7 99.6 ± 0.1 94.8 ± 0.9 93.5 ± 0.3 97.5 ± 0.1 95.5 ± 0.2

Macro avg 83.6 ± 0.4 82.9 ± 0.5 82.6 ± 0.5 84.6 ± 0.3 83.7 ± 0.6 83.5 ± 0.6 79.7 ± 0.1 80.1 ± 0.1 79.4 ± 0.1

Notes. Classification precision, recall, and F1 score per class and macro average of the models put into production. We include the two best-performing ATAT variants (with MTA) and the RF-based baseline.

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.